Bug 146818

Summary: Error when multiplying, dividing, subtracting or adding two Floating Point fields
Product: [Applications] KEXI Reporter: r.c.drew
Component: QueriesAssignee: Jarosław Staniek <staniek>
Status: CLOSED FIXED    
Severity: normal    
Priority: NOR    
Version: 1.1.3 (KOffice 1.6.3)   
Target Milestone: ---   
Platform: Ubuntu   
OS: Linux   
Latest Commit: Version Fixed In:
Sentry Crash Report:

Description r.c.drew 2007-06-15 12:16:25 UTC 
Version:           1.1.3 (using KDE KDE 3.5.6)
Installed from:    Ubuntu Packages
OS:                Linux

In the following: FP = Floating Point, Int = Integer
If you set up a query to multiply FP x Int, result is FP.  This works fine.
However, the result for FP x Int gives a square with a question mark inside.
Whether the FP number is single or double precision, the same problem occurs.
This has been confirmed by Jarosław Staniek (current project maintainer) who asked me to file the bug.
The same problem exists when adding, subtracting or dividing two FP numbers.
Clearly, Kexi has great potential, and much great work has been done, but it is unusable as a database if any FP numbers are involved!
Comment 1 r.c.drew 2007-06-15 12:28:38 UTC 
Sorry, I made a typo: I should have written "However, the result for FP x FP gives a square with a question mark inside".
Comment 2 Jarosław Staniek 2008-01-10 17:41:01 UTC 
SVN commit 759457 by staniek:

KexiDB

fix the bug #146818: handling expressions containing 
binary operators like *, /, -, + where both arguments are of 
floating-point type (with single or double precision). 
Previously the result was marked as boolean, now it is 
floating-point.

CCMAIL:146818-done@bugs.kde.org
CCMAIL:kexi@kde.org



 M  +2 -6      expression.cpp  


--- branches/koffice/1.6/koffice/kexi/kexidb/expression.cpp #759456:759457
@@ -400,14 +400,10 @@
 	if (ltInt && rtInt)
 		return KexiDB::maximumForIntegerTypes(lt, rt);
 
-	if (Field::isFPNumericType(lt) && rtInt)
+	if (Field::isFPNumericType(lt) && (rtInt || lt==rt))
 		return lt;
-	if (Field::isFPNumericType(rt) && ltInt)
+	if (Field::isFPNumericType(rt) && (ltInt || lt==rt))
 		return rt;
-	if ((lt==Field::Double || lt==Field::Float) && rtInt)
-		return lt;
-	if ((rt==Field::Double || rt==Field::Float) && ltInt)
-		return rt;
 
 	return Field::Boolean;
 }